Finite Element Model Size Reduction Research and Validation Study

In the past years, the scope of finite element models (FEM) has grown significantly. FEMs can be used to analyze various types of dams, with processing time varying based on the complexity of the model. FEMs used to inform dam safety decisions now contain over two million elements, with the potential to increase depending upon project complexity. These models are generally used to assess the likelihoods of spillway-related seismic potential failure modes (PFM) and incorporate features such as the spillway, dam, reservoir, and surrounding landscape.

Dynamic analysis run times can take as long as one day per second of ground motion time history. This means that running a simulation of a 21-second-long earthquake can require as much as three weeks of computing time. If a dam should happen to be located near a subduction zone area, in which ground motions can last as long as 300 seconds, computations may require months to run. Such complex models are time-consuming in their initial mesh generation, and as run times protract, they can extend project schedules in the latter stages as well. In an effort to deliver the greatest value to every client, this research was undertaken to investigate and test smaller FEMs and asses their quality and accuracy. Such efforts could have the potential to reduce cost to the client.